Kubernetes has become the standard way, and many will prove it by deploying containerized applications at different scales. But if Kubernetes helps us deal with messy and complex container deliveries, what will help us deal with Kubernetes? It can also be complex, confusing and difficult to manage.
As Kubernetes grows and develops, many of its nuances will, of course, be eliminated within the project itself. But some users don't want to wait until Kubernetes gets easier, so they've developed their own solutions to many common Kubernetes production problems.
NB I hope that the unknown bat infection that bit the dog, which bit the pangolin, which bit the Chinese by a strange coincidence in Wuhan, where the BSL-4 level biological laboratory is located, will subside by February and we will only remember 2019-nCoV using obscene vocabulary . And we can do it offline February 8β10, 2021, and for advanced users of K8s February 12-14. Honestly, personally, as an editor, I missed the drive, coffee breaks, disputes and tricky questions from the speakers. Well, or the whole planet will die out in the style of the most cruel and trashy novels by Styopa of our Korolev, if the powers that be are tired of our dumb jokes like Conchita Wurst, Patriarch Kirill's watch and the desire of the Pope to correct the words of the Lord's Prayer.
But back to the main point.
Goldpinger: Visualizing Kubernetes Clusters
People prefer to watch. Graphs and charts make it easier to understand the big picture. And given the scale and complexity of the Kubernetes cluster, we can use this feature to its fullest.
A project with a funny name (probably something about agent 007, approx. translator) , an open source tool published by Bloomberg Technical, is a simple tool that runs inside a Kubernetes cluster and displays an interactive map of relationships between nodes. Normally functioning nodes are shown in green, inoperable ones in red. Just click on a node to find out the details. You can also customize the API with Swagger to add additional reports, features, and more.
K9s: Full screen console interface to Kubernetes
Sysadmins love "single-window" goodies. is a full-screen console interface for Kubernetes clusters. With it, you can view running Pods, logs, and deployments easily and naturally with quick shell access. Note that you need to grant read user level and namespace permissions to Kubernetes users for K9s to work properly.
Kops: Console ops for Kubernetes clusters
development from the Kubernetes team will help you manage Kubernetes clusters from the command line. It supports clusters running on AWS and GKE and also works with VMware vSphere and other environments. In addition to automating the install and uninstall process, Kops can help you manage other types of automation as well. In particular, he can create settings for Terraform, which can be used to reload the cluster using Terraform.
Kubebox: A terminal shell for Kubernetes
Advanced terminal shell for Kubernetes, , does more than the good old Kubernetes wrapper and API. Among other things, it can show in real time the use of processor time and RAM, a list of pods, the contents of logs, and also launch the settings editor. What I also like is that it is available as a standalone application for Linux, Windows and MacOS.
kube-applier
installs as a Kubernetes service, gets the declarative Kubernetes cluster settings from the git repository, and then applies them to the pods in the cluster. Each time changes are made, they are taken from the repository and applied to the requested pods. This is somewhat similar to Google's Scaffold, but works to manage an entire cluster instead of a single application.
It is possible to make changes to the settings on a schedule or upon request. All actions are logged, and Prometheus-compatible characteristics are also presented, so you can always see what might affect the behavior of the cluster.
Kube-ps1: Smart command line prompt for Kubernetes
No, it's not a Sony PlayStation emulator for Kubernetes, although that would be neat. This is a simple Bash command line extension that displays the current Kubernetes context and namespace in a tooltip. Kube-shell includes it in many other features, but if just a smart hint is enough for you, Kube-ps1 will provide it for you at minimal cost.
kube-prompt
Another minimal, but very pleasant to use Kubernetes CLI modification is , which allows you to enter an interactive session with the Kubernetes client. kube-prompt saves you from having to type kubectl before each command, and also provides auto-completion with contextual information for each command.
Kubespy: Real-time monitoring of Kubernetes resources
by Pulumi is a diagnostic tool that helps you debug changes to a cluster resource in real time by providing a sort of text panel to control what's going on. For example, you pod states since startup: the pod definition is written to etcd, the pod is scheduled to run on the node, the kubelet on the node creates the pod, and finally the pod is marked as running. Kubespy can be run as a standalone program or as an extension to kubectl.
Kubeval: Checking Kubernetes settings
Kubernetes configuration YAML files can be human-readable, but that doesn't always mean they can be verified in the same way. It's easy to miss a comma or a name and not find it until it's too late. Better to use , installed locally or connected in the CICD pipeline. Kubeval takes a YAML definition of Kubernetes settings and feeds back correctness information. It can also output data in JSON or TAP, and parse source templates referenced by Helm chart settings without making additional queries.
kube-ops-view: panel for multiple Kubernetes clusters
Kubernetes already has a pretty good general purpose dashboard, but the Kubernetes community is experimenting with other ways to display data useful to Kubernetes sysadmins. there is just such an experiment, it provides an opportunity to review several clusters, you can see the consumption of processor time and RAM, the state of the cluster modules. Please note that commands cannot be called, the tool is only for visualization. But the displays it provides are crisp and smooth, begging for a wallboard in your support center.
Rio: Delivering Applications for Kubernetes
, a project from Rancher Labs, implements common Kubernetes application delivery practices such as CD from Git, AB, or blue-green distributions. It can also roll out a new version of your application as soon as you commit the changes, helping you manage complexities with e.g. DNS, HTTPS, Service Mesh.
Stern and Kubetail: Viewing Logs in Kubernetes
produces colored output (as the command tail) from pods and containers to Kubernetes. It is also the fastest way to get the output of multiple sources into a single stream that can be read on the fly. At the same time, you have a visually distinguishable way (by color) to separate the streams.
merges logs from different pods into one stream in a similar way, color-coding different pods and containers. But Kubetail is a Bash script. so it doesn't require anything else but a shell to run.
Only registered users can participate in the survey. , you are welcome.
And what do you use to simplify routine operations?
-
Present in several = 2,9%Goldpinger1
-
Present in several = 22,9%K9s8
-
Present in several = 0,0%Kops0
-
Present in several = 0,0%Kubebox0
-
Present in several = 0,0%kube-applier0
-
Present in several = 0,0%Kube-ps10
-
Present in several = 0,0%kube-prompt0
-
Present in several = 0,0%Kubespy0
-
Present in several = 2,9%Kubeval1
-
Present in several = 0,0%kube-ops-view0
-
Present in several = 0,0%Rio0
-
Present in several = 2,9%Star1
-
Present in several = 5,7%Kubetail2
-
Present in several = 28,6%None of this10
-
Present in several = 5,7%I have my own "pree-e-e-le-e-essst"
-
Present in several = 8,6%I'll try to feel something from the list 3
-
Present in several = 20,0%I control Kubernetes with a neuroimplant, like in the movie "Johnny Mnemonic"7
35 users voted. 19 users abstained.
Source: habr.com